Research On Identification Method Of Passenger Flow Congestion State In Intervals Of Urban Rail Transit

With the rapid increase of rail transit travel sharing rate,the contradiction between capacity and demand is becoming more and more prominent,and the phenomenon of station flow restriction,crowded passenger flow and overloaded train rate occurs frequently,which not only reduces the level of rail transit operation service,but also makes the operation safety risk higher.In this context,how to scientifically and reasonably grasp the crowded state of rail transit network becomes the key to improve the efficiency of rail transit operation and ensure operational safety.Accordingly,a theoretical and computational method of using the line interval congestion index to describe the line interval congestion status is proposed,and the urban rail transit congestion status identification model is established.The specific research contents and conclusions are as follows:(1)On the basis of the definition of crowded passenger flow,the spatial and temporal characteristics and propagation characteristics of rail transit crowded passenger flow are analyzed,and the main influencing factors of crowded passenger flow are summarized and formed.The "station congestion index" and "cross-sectional passenger flow" are used as the key indicators of the line interval congestion index.The model for calculating the interline congestion index of rail transit lines is constructed.(2)The method of estimating the cross-sectional passenger flow as the key index of the line interval congestion index is studied.This method mainly includes the following elements: searching the effective paths between any OD pairs of rail traffic according to the improved shortest path algorithm,assigning the passenger flow to each path using Gaussian mixture model(GMM),matching passengers to each train according to the relationship between passenger entry and exit times and train schedules,and calculating statistics to derive the cross-sectional passenger flow.(3)Based on the calculation of the line interval passenger flow congestion index,the ACO-FCM improved clustering algorithm is proposed by combining the advantages of the fuzzy clustering algorithm(FCM)and the ant colony algorithm(ACO)to identify the line interval passenger flow congestion state.The problems of slow convergence and easy generation of local optimal solutions of fuzzy clustering algorithm are solved.(4)The line interval congestion index calculation model and the ACO-FCM improved clustering algorithm are applied to the Chongqing rail transit network,and the results of line interval congestion index and passenger congestion state identification are output.Combined with ANOVA analysis of variance,the clustering effects of fuzzy clustering algorithm and ACO-FCM improved clustering algorithm were compared and analyzed.The results show that the mean variance of the ACO-FCM improved clustering algorithm is larger than that of the fuzzy clustering algorithm,which is 185.6% of the mean variance of the different levels of passenger congestion indices.The mean squared difference between passenger congestion indices within the same class is smaller,which is 15.3% of that of the fuzzy clustering algorithm.The above shows that the clustering effect of ACO-FCM improved clustering algorithm is better and verifies the effectiveness of this passenger flow congestion state identification method.Based on the AFC data of urban rail transit line interval crowding index model and the improved line interval passenger flow crowding status identification algorithm,the evaluation index system of line interval passenger flow crowding status is improved,the accuracy of urban rail transit line interval passenger flow crowding status identification is effectively improved,and the problem of fuzzy passenger flow crowding status classification is preliminarily solved,It provides a reference basis for the urban rail transportation management department to formulate the crowded passenger flow decongestion plan.